Thank you for open-sourcing MergeBench and providing such a comprehensive evaluation suite for model merging!
I am currently looking into the generalization and forgetting aspects of merged models. While the repository contains the code for the main in-domain evaluation tasks, I couldn't find the specific evaluation scripts or configurations used for the generalization benchmarks.
Could you please provide the evaluation scripts or point me to the implementation details for the following datasets mentioned in the paper?
Thank you for your time and for contributing this great resource to the community!
Thank you for open-sourcing MergeBench and providing such a comprehensive evaluation suite for model merging!
I am currently looking into the generalization and forgetting aspects of merged models. While the repository contains the code for the main in-domain evaluation tasks, I couldn't find the specific evaluation scripts or configurations used for the generalization benchmarks.
Could you please provide the evaluation scripts or point me to the implementation details for the following datasets mentioned in the paper?
Thank you for your time and for contributing this great resource to the community!